The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill provides endpoints designed to receive and process ElevenLabs webhook events. This ingestion of untrusted external data constitutes a surface for indirect prompt injection if the event data is subsequently used in agent prompts without adequate sanitization.
Ingestion points: Webhook handler routes located in examples/nextjs/app/webhooks/elevenlabs/route.ts, examples/express/src/index.js, and examples/fastapi/main.py.
Boundary markers: All implementation examples include HMAC-SHA256 signature verification (using either the official SDK or a manual implementation) to ensure the authenticity of the incoming data.
Capability inventory: The example code primarily logs event types and data to the console, providing placeholders for user-defined business logic.
Sanitization: Authenticity is verified via cryptographic signatures before the payload is parsed and handled.
[EXTERNAL_DOWNLOADS]: The skill references and utilizes external packages and tools, including the official @elevenlabs/elevenlabs-js SDK and the hookdeck-cli for development environments.

elevenlabs-webhooks